A friend tried Rekkon a few weeks ago and described the difference like this: "I like how you can just talk to it. With MyFitnessPal you have to be so specific."
She wasn't talking about a technical feature. She was describing a completely different relationship with the act of logging food.
What "being specific" actually costs
When you open MyFitnessPal to log lunch, you type what you ate. The app searches a database. You scroll through results. You find the closest match. You adjust the serving size. You confirm. Repeat for every component of the meal.
For a simple lunch — say, a chicken wrap from the local café — that's probably four or five separate database searches, portion adjustments, and confirmations. It takes two to four minutes. Every day. For every meal.
Two to four minutes sounds small. But multiply it by three meals, 365 days, and you've spent 36 to 73 hours a year doing data entry.
"Logging food seems to have been made by and for people with OCD." — App Store review of a leading calorie tracker, 2025
The specificity problem isn't about accuracy
You might think: all that detail makes for better accuracy. But that's not what the research shows. Studies consistently find that even careful calorie counters underestimate their intake by 20–30%. The precision of the database entry doesn't translate into precision of the calorie count, because the food itself is variable. A "medium chicken breast" from one source might be 150 grams; from another, 220 grams. The database entry and the actual food are not the same thing.
What the research does show is that consistency matters far more than precision. Someone who logs directionally every day for six months will get better results than someone who logs precisely for three weeks and burns out.
On the accuracy of rough estimates
A meta-analysis of portion estimation methods found that hand-based estimates fall within 10–20% of actual weights for most common foods. For the purposes of weight management — where day-to-day trends matter more than individual meal precision — this level of accuracy is more than sufficient.
What voice changes
When you can say "three slices of leftover pizza and a banana smoothie" and have that logged in five seconds, the cognitive load drops. You stop dreading logging. You do it automatically, the way you might send a voice message instead of typing one.
The other thing voice changes is the conversation. Once logging is just talking, asking questions feels natural too. "I'm going out for dinner tonight — any tips?" You've opened a conversation with something that knows what you've eaten today and can give you a real answer.
The voice revolution in nutrition apps
Voice logging isn't entirely new — MyFitnessPal added a voice logging feature to its Premium tier in 2025, and several smaller apps have offered it for years. But most implementations still require the app to match your spoken words against a database, which reintroduces the search-and-select friction in audio form.
The difference with AI-native voice logging is that the app understands context, not just keywords. "What I had at the barbecue on Saturday" is a different kind of input than "200g chicken breast, grilled." Both describe the same thing. Only one of them is how people actually eat and talk about food.
Common questions
Can you log food by speaking instead of typing?
Yes. Voice-first nutrition apps like Rekkon let you hold a button, say what you ate, and have calories and macros estimated automatically. The logging takes about five seconds and works hands-free — useful while cooking, driving, or at the gym.
How accurate is voice food logging?
Voice logging using AI description analysis falls within 10 to 20 percent of actual calorie content for most common foods — comparable to database entry accuracy, which also relies on user estimates of portion size. For weight management, this level of accuracy is sufficient.
Why is MyFitnessPal harder to use than talking to an app?
MyFitnessPal requires searching a 14 million entry database, selecting the closest match, adjusting serving size, and confirming — a process that takes 2 to 4 minutes per meal. Voice logging replaces this with a five second spoken description, dramatically reducing friction and improving consistency.
Seven days. See what it notices.
No barcodes. No databases. No meal plans. Just talk to it.
Try free